Accelerate business outcomes with AI agents and Onepoint’s AI Ecosystem for Boomi

Linda Lam

Hi, everyone. Thanks for joining. I’m just going to give it a couple more minutes for attendees to just trickle in, and then we’ll start.

Okay. Let’s kick this off. All right. Hi, everyone. Welcome to today’s webinar session titled Accelerate Business Outcomes with AI Agents. Thank you so much for joining us today. I am genuinely excited to host this session because we have one of our fantastic tech partners, Onepoint, with us today.

They actually presented earlier this year at Boomi World in Dallas, and the feedback from that session was so positive and so strong. So we wanted to bring their great work and showcase it to a broader audience. So that’s why we’re here today. Today’s session is really their show. Over the past several months, the Onepoint team has invested significant time, effort, energy into building powerful AI agents, both within Boomi Agent Studio as well as other areas of services on our Boomi platform. So this is really a great opportunity for them to showcase the innovative agents they develop and the business impact these agents are delivering to customers.

Before we dive in, let me just quickly introduce myself. My name is Linda Lam. I lead the Technology Partner Program here at Boomi. I’ll be your moderator for today’s session. Joining us from the Onepoint team, who are going to be the main presenters today, we have Rafiq, who is the Business Development, and Miguel, who’s the Principal Solutions Architect. Thank you both for joining us today.

Okay, before we dive into the session, I’ll just quickly go over the agenda items. Rafiq will start with a brief overview of the Onepoint organisation. Then Miguel will jump in and give you a snapshot look at the incredible agents that they have developed and built, now available in our Boomi marketplace.

Get ready because the work that they’ve accomplished here is truly impressive. So I’m really excited for you all to see that. Then we’ll jump into the demo, which I’m also excited for you all to see because Onepoint team will play out real business scenarios and show how their agents come to life and solve key business challenges.

And then finally we’ll leave time for your questions if you have any questions at any point during the presentation please drop it in the Q&A box um if we don’t get to it at the end of the session don’t worry the team here will follow up with you all um so without further ado i’m gonna hand it over to the Onepoint team.

Rafiq Rajabally

Hi everybody thanks very much for joining us appreciate, you’re waiting while we’re getting things ready. So what I’m here to do is give you a quick overview about Onepoint, which is probably a company that you’ve not come across before. And there’s a lot to tell, but I’m just going to keep it to the headline.

So this is our 20th year of trading. We’ve been five years a partner with Boomi. If I was going to point to a USP, I would say it’s that we are very much enterprise systems, architecture, DNA company. Okay, so that’s at the root of it. And then from there, we’ve been able to do data better and innovate with AI, which we’ll come on and bring to life later in the presentation.

We are a 60-person company with offices in London, Manchester and Pune. And we have been fortunate to work with some leading brands. So we’ve got, I think the key thing to stress here is that we have deep industry expertise across manufacturing and aerospace, retail, property, management or real estate, I think you call it in the States, and energy.

So we’ve got a good, solid cross-industry footprint that is international. We’ve worked with some of the biggest names in business from FTSE 100 and Fortune 500 companies to Global 1000 Enterprises. And, you know, we are a values-driven business. And what that means is that we believe in high integrity ethical practices and giving back and being successful has enabled to give us to give back to charities and to the local community and we have also developed apprenticeship programs for people that want to get into the industry so enough of that let me introduce Miguel who’s going to take you through Onepoint’s AI Ecosystem for Boomi. Miguel.

Miguel Vale

Thank you so much, Rafiq, for all the welcome. I will present to you what we have been working so hard in the last few months, which is the Onepoint AI Ecosystem for Boomi. So the AI ecosystem for Boomi is divided in three parts.

The first is this set of connectors that we have built, which is the nuts and bolts, and these nuts and bolts will be able to create agentic workflows. What is an agentic workflow? It’s a smart process that can make decisions and can be wrapped up as a web service. Those web service can be used in AI agents. Basically, the majority of the agents that we have built, we have used the Boomi Agent Studio, except one of them that we use that flow application.

But this AI Gen Studio, you can basically code it and a standard API that the LLM will deal with it, but also you can plug in agentic workflows, which is basically smart APIs that can make decisions.

Let’s jump now to the connectors. So we have built a few connectors. The most important one is the OP Intelligence, which is renamed to OP AI LLM Intelligence. It’s basically a wrapper for any LLM.

Basically, you code your process using, for example, OpenAI, but tomorrow you can change to another provider, for example, Anthropic or Gemini or Llama or anything else. You don’t need to change your process, you just need to change your connection properties and voila. That connector also supports memory, and we are working to add MCP functionality to that connector.

We are also building another connector, what we call OP-AI agent-to-agent. And this basically is a connector that allows agents to communicate with other agents, offloading tasks to agents, checking status, informing the feedback for agents.

Another connector that we have is the OP database with RAG. So it’s basically a spin-off from the database v2 connector from Boomi, which we improved and adding support to other data types like enums, arrays, and so on, and also vectors and similar search. It supports the majority of the database that you already know, like for example, Oracle, Microsoft SQL Server, MySQL, Postgres, IBM DB2, and more.

We have also the OP Web Search connector that you can select the provider, you can select Bing or Google or another provider and basically to pull fresh data to your processes. We have finally the OP Wikipedia, which again also pulls fresh data from Wikipedia.

Now, here you see 10 agent workflows and basically all these 10 are available in the Boomi marketplace as the recipes so you can do you can go download check how they are built change it test it I will not go through all of them I will just like a couple of them just for you to get to have the gist of it for example the PII data detector. It’s basically a process that inspects payload data. If it detects any type of PII, Personal Identifier Information, could be financial information or health information, you can act on it. For example, you can throw an exception or you can reject or you can log to the database. You can do whatever you want.

Miguel Vale

Another example is the hotel matcher. This comes from a need from a client of us that need to standardise hotel data. So it would receive text hotel data and you want to identify which is the entry in the database.

So we end up creating a RAG system with millions of hotels and with the similar search, we find the top more prominent hotels sent to LLM and identify the right one to standardise. Now we have here our blueprint of the 27 AI agents that we have built.

Again, this is not a eye test, so you don’t need to understand everything that is here. I will just highlight the most important things. They are divided in brands and these brands are very aligned with Boomi and with ourselves. For example, AWS, Azure, SAP, ServiceNow and others.

You can see as well that some has a colored magenta dots. This means that they use agentic workflow. They use an API that is smart, that can make decisions. I will just highlight a few of them, not just the two or three. For example, the Boomi Process Monitor, the agent workflow basically downloads the logs from Boomi, which is an asynchronous process and search on those logs. The DataHub AI agent Quarantine Resolver that is also, has also an agentic workflow, which can not just detect quarantines, but can also resolve some types of the quarantine issues automatically.

And for Onepoint, we developed a lot of them. I will not go through all of them. Just one is an example. The database explorer, it’s an AI agent that you connect to any database, that’s it. And you can just query the database using natural language. You can use English or Spanish. It will identify the relevant tables, the relevant fields. You can query information, and even you can update and change information on that database.

And I think we are ready for our scenarios. So I will hand over to Rafiq.

Rafiq Rajabally

Thanks, Miguel. So what we’re going to do here is paint a picture using our AI agents. So in this scenario, the first scenario, it’s an online portal company that has launched a promotion. And this has triggered a spike in website traffic, but the website’s not up to it for the moment. We don’t know why. This has caused a number, a huge surge of calls into the contact centre. People that are unable to actually get on the website or get in touch with the contact centre are now going to social media, so it’s not a great picture, and they’re starting to complain quite vociferously.

In this scenario, I play Rory. I’m VP Customer Services for Online Portal Co. And Malcolm is a Client Services Director, looks after our account from a managed service perspective.

Miguel Vale

And I will present you four AI agents that we have built, the Boomi Process Monitoring to understand what’s going on now, the Eyer Anomaly Detection to predict features, issues, the AWS Resource Spinner that will allow us to spin new resources, and the ServiceNow Incident Management to update incidents in ServiceNow.

Let’s start.

Hi, Rory. What’s up? How are you?

Rafiq Rajabally

Malcolm, well, I could be better. You know, we’ve got this promotion running. The website is not dealing with it well. We’re getting inundated with complaints on the contact center. That’s blown up. And we’re starting to see very negative stuff coming on social media.

So I’m not good at all, especially for the promotion. The other thing is I’ve got an industry dinner or something tonight which I need to go to. I would very much prefer the subject not to be our website crash. What can you do to help?

Miguel Vale

Rory, okay, so first of all, relax, okay? We have a pretty set of AI agents that can help us today. And maybe you will be alleviated from this pressure and go to your dinner quickly. So shall I show you some of the AI agents that can help us today? Yeah, let’s jump on it. Okay, so let me share to you.

The first one is the Boomi Process Monitoring and we will try to understand what’s going on in our website. So let me ask information about the five last executions. Okay, so it’s very simple. What’s going on with the last five executions? And our AI agent said that it’s not great, it takes about 30 seconds where we typically respond in less than a second. So something is fishy in our website, Rory.

Rafiq Rajabally

Yeah, that’s really poor. There’s something definitely wrong.

Miguel Vale

Yeah, so let me show you another AI agent that identifies if this is a blip and we don’t need to worry about it or if something is more serious, okay? So this is the Eyer Anomaly Detection.

So this agent analyse what’s going on now, analyze the future and will predict some forecast for us. Okay. So basically the forecast is not great. So the peak time it will be at 2:00 PM, which is in a few minutes time and it can reach 40 seconds delay time. So it’s not really great news, Rory. So we need to do something.

Rafiq Rajabally

Yeah, we need to get on it. So what do you suggest? Where do we go from here?

Miguel Vale

So we have these new brand new powerful AWS Resource Spinner. I think this agent can help us. I will just give you a glimpse of it. So basically what he can do basically is

It can spin EC2 instance, it can spin S3 buckets, it can spin databases, it can spin VPCs and so on and so on. But what’s really important here is to spin a boomy runtime environment and join the cluster to give some breathing room to our production environment.

Rafiq Rajabally

Do we really have time for that? I mean, that sounds like it’s a lot of work and it’s going to take ages.

Miguel Vale

I know what you think, Rory. In the past, it would take a couple of days just to spin a new Boomi runtime. But with AI agents, this is done in minutes.

Let me show you. I want to show you. Let me just ask the AI agent to deploy a Boomi runtime on the default VPC. Let’s see what he will do. He’s configuring and he’s setting up the Boomi runtime environment with auto-restarts, so we don’t need to worry about something was wrong with that item. He’s ready. We already have an IP and it’s ready. Let me show you on the Boomi platform.

So if you see in the Boomi platform, it’s blue and it’s ready to join the production cluster. So I just need to drag and in a few seconds it will be ready. So you can go to your dinner, Rory.

Rafiq Rajabally

So you’re telling me that this is live now?

Miguel Vale

Yeah, it will be live in a few seconds. Okay, don’t worry. Okay. Yeah, yeah. And I will keep an eye on it. I know that I will keep an eye on it for the in the next few days as well.

Rafiq Rajabally

Okay. And I raised the service now ticket. I’m not sure if I mentioned that. I don’t know if you’ve seen it. But can you can you deal with that for me? Because I need to get away.

Miguel Vale

Yeah, yeah. Okay, let’s let’s ask to another AI agent for that. Okay. And if you see, so let me connect to the ServiceNow Incident Management. And let’s just update that ticket saying that we have seen another Boomi runtime. And we will monitor for a few days. And voila, done. So the incident was resolved. And I will show you in the ServiceNow UI that it’s being updated, has that comment and the state is resolved. So I think it’s done, Rory. You can go for your special dinner.

Rafiq Rajabally

Okay, I’m relieved. Thank you, Malcolm.

Okay, so in terms of scenario one, let’s consider the business outcomes. Again, fulfilled by Boomi agents. So firstly, we’ve identified a bottleneck. So the scenario that we’ve described, think about Amazon, think about Shopify, where they’ve had outages, how much time it takes just to understand what’s going on.

Secondly, we’ve used the AI agent, which has looked at the history of what’s not worked. It’s extrapolated that data to predict what could happen in the future.

It wasn’t looking very rosy, so it’s recommended that we spin up another resource. We’ve actually taken you through the process that we would use to spin up another environment. So we’ve instantly scaled and it is sort of business as usual. And I think the key thing here is when you consider the work that’s required to spin infrastructure, to get that running, to do the installation and configuration, to do the verification and testing and all the stuff that goes with that. I mean, you’ve condensed a day, two days work into minutes and that’s a significant advantage for business.

Miguel Vale

Thank you so much Rafiq. I think we are ready for scenario two.

Rafiq Rajabally

Okay, scenario two. In this scenario, it’s an e-bike rental company that has struck up deals with hotel chains in the US to be the e-bike vendor of choice. So what that means is that they’re going to be hopefully a lot of e-bikes on the road and we’re going to need the spares to keep these bikes on the road because you know the reality is nobody really cares for these bikes. We get things like seats broken, tires which need to be replaced, baskets which are damaged.

And in this scenario, I’m Ralph, I’m the head of partnerships. Mike is our supply chain manager. And we’re going to put the problem to an AI agency, what it can do to help us.

Miguel Vale

And I will present you three AI agents that we have built. The SAP Low Stock Purchasing Recommendation AI agent, the Data Hub Golden Record Explorer, and the U.S. Government Search that connects to government websites to check in real-time information.

Okay, so hi, Ralph. What’s up? Okay, I heard this great news. I heard that great news.

Rafiq Rajabally

Yeah, we’ve got the deal done. So we’ve got the hotel chains on board. We are the e-bike partner of choice. Now we need to make sure that we can keep these bikes running. And as you know, you know, they’re items on there that get either damaged or need to be replaced. So I’m thinking about seats, I’m thinking about tires, and I’m thinking about baskets. So what can we do to stay ahead of this situation, Mike?

Miguel Vale

So, Ralph, I have really, really great news for you. We have a bunch of AI agents that can resolve all these things in no time. Let me show you. So I will jump to one AI agent that is our brand new SAP Low Stock Purchase Agent. And I will just tell me what about the stock recommendation for these materials that you just enumerated, tires, seats, and baskets.

Miguel Vale

So, it will connect to SAP, it will analyse vendors, it will analyse stock quantities, and the recommendation is simple. It basically says that we should hire 25 tires from Michelin today. For seats, 30 units are required to be ordered today, and for baskets, no need; we have plenty of stock.

Rafiq Rajabally

Okay, that’s interesting that it mentioned Michelin.

Rafiq Rajabally

I was reading that Michelin have production issues. These production issues are causing extended delays and subsequent delays to fulfillment. What is it that we can do about that?

Miguel Vale

Yeah, I heard that Michelin is having a weak delay. So let’s add that information to our AI agent and let’s see what it recommends.

So I’m just asking, I’m just saying that the vendor Michelin North America is expecting a delay of a week. Can you please factor this in the recommendation for the materials of tires? And he will again connect to SAP, again checking vendors. And basically he recommended us to split and order from Goodyear. Okay. Same units, 25 units today.

Rafiq Rajabally

Okay, okay, that’s interesting. Are Goodyear someone that we’ve done business with? Are they on our books?

Miguel Vale

I think they are, but it was long time ago. But let’s check information with another agent. This is the Golden Record AI agent that we can ask if Goodyear is still an active supplier. Just give me details of Goodyear. Is it active? Our Golden Record AI agent will tell us that yes, he’s active. Do you need some contact details?

Rafiq Rajabally

That’ll be useful. Yeah, what have you got?

Miguel Vale

Okay, so let’s ask the agent to give me some more complete details. And it’s connecting to the data. Okay, pulling those information. And yes, we have phone number, contact, email, and even address if you want to pop in. Anything else, Mike?

Rafiq Rajabally

Yeah, I always like to be sure because you know, you’ve got businesses with similar names. Can you check that Goodyear is still trading. Is there a way that we can go to sort of an official site just to get verification of that?

Miguel Vale

So we have a US Company Search AI agent that connects to the government website. So let me spin that and let’s ask if Goodyear is still a registered company in the US and let’s see what this agent says.

Miguel Vale

So he’s retrieving information and yes, he’s confirmed that Goodyear is really registered and we can do business with them.

Rafiq Rajabally

That’s good to hear. Okay. You wouldn’t have an AI agent that could actually place the order for me?

Miguel Vale

I know what you’re thinking. We are in the fact we are building that functionality in our SAP Low Stock Purchasing, but it’s not really quite ready. So, Ralf, I think you need to grab the phone this time.

Rafiq Rajabally

Yeah, phone or email. I thought you were going to say that. Okay, thank you.

Miguel Vale

Okay, thank you so much.

Rafiq Rajabally

So in terms of outcomes, I think when you’re thinking about contingencies for supply chain, they can take weeks, it can take months, depending on the complexity of the supply chain. What we’ve demonstrated here with our agents is firstly, we’ve found an alternative supplier and we’ve been able to verify details about it.

We’ve used the measure of intelligent sourcing to come up with recommendations and now we’re ready to go to meet that demand. So that’s all very exciting stuff and it was done literally in minutes using our Boomi agents.

Miguel Vale

And we are ready for the next scenario, scenario three.

Rafiq Rajabally

Okay, in scenario three, the company is Perfect Fit Company. It’s a luxury bespoke clothing manufacturer that designs articles for either corporates or for individuals. The key issue here is that they’ve made an acquisition of two complementary businesses.

The inconsistent product data that they have across all their brands means that customers are more often than not disappointed, if not unhappy. They’re demanding refunds or credits. And again, with so much going on with social media, the reputational damage can be immense and we need to staunch that.

I’m Ryan, I’m the VP of Customer Service. I’m talking to our Chief Data Officer, Max.

Miguel Vale

And I will present to you the Data Management Smart Advisor AI agent. This Smart Advisor AI agent was built with the Flow application and the agentic workflows.

Miguel Vale

It is not using the Boomi Studio. And basically it’s a self-service tool. Tell us what we don’t know. And in terms of cost, it costs a fraction of the cost of a consultant. And how this smart advisor works?

We don’t need to worry about asking questions to him. He asked questions to us, try to understand what’s going on, build a confidence level. When he understands fully what’s going on and the confidence level is high, he will give us a recommendation, what to do, what not to do, and also the expected outcomes.

Let’s start the demo. Hi, Ryan. What’s up?

Rafiq Rajabally

Max, well, good news. We’ve completed the acquisition of those two companies.

Miguel Vale

Congrats!

Rafiq Rajabally

It’s team effort. I think we’re all happy. But like all things, there are challenges and we need to get to the root of those challenges and address them, which is why I’m talking to you. We’ve got some data issues and we need to standardise right across all these three businesses.

Miguel Vale

Right, Ryan. So we have this new self-service tool, which is a Smart Advisor AI agent that I think can help us because really we don’t know how to start. And this AI agent can help us to really understand how to start. Okay. So shall we jump on it?

Rafiq Rajabally

Yeah, let’s do that. And you’re saying it’s self-service, right?

Miguel Vale

Yeah, self-service, it won’t cost you thousands of pounds. It’s really cheap. For us, it almost doesn’t cost anything.

Now, as you can see, there is an initial question and there is potential answers. Let’s ignore for now the answers and let’s focus the question. And I will guide you to put the right information. So basically what today the smart advisor wants to know is what concerns more? So we had acquired these two complementary companies, but what are the most concerns for us, Ryan?

Rafiq Rajabally

Aside of sort of the two companies which you’re correct to point out historically we’ve always suffered from data issues and I would say they fall into two areas there’s the sort of missing technical specs on the garments.

And another example would be, let me see, inconsistent color representation across the ranges. So what’s blue for one set of items is sky in another, and we need to get that straight. And I think it just underlies this sort of theme that goes across the business, which is we really do need to get our data house in order and standardise it.

Miguel Vale

Yeah, okay, I got it. Okay, so let’s just…

Set up that information and let’s send to our Smart Advisor. Okay. And that Smart Advisor now we will connect to a RAG system data management. It will update the confidence level. It will understand if he knows enough or not enough. And it will ask more questions until he reached the confidence level high. And then he can provide us information.

Okay. Okay, voila. Now he’s asking the second question. And the second question is that he wants to know a specific challenge. So, Ryan, tell me, what are the challenges that we are facing here?

Rafiq Rajabally

Yeah, that’s a good point. Well, even though we’ve made those acquisitions, we’re still a small company. And we’re not techies. Yeah, we don’t have any technical staff.

Miguel Vale

Okay, yeah, it’s a great challenge. Okay, anything else?

Rafiq Rajabally

Yeah, I would also say that we’re spending more time giving money back, making refunds and credits than we are actually selling. And that’s not a tenable position to be in.

Miguel Vale

Yeah, so we are not doing the business activities that we need to do. Let’s add that information to our smart agent, send that to our smart agent. Again, it will connect to the RAG system, will do a similar search, will connect to LLM, will update the confidence level, and it will add probably one more question.

Miguel Vale

Okay. And now the question that he’s asking is, he wants for us to be specific. Can you give me a couple of specific examples, Ryan?

Rafiq Rajabally

Okay. So one of the companies that we’ve acquired is based in the US. And as you know, American sizing, the measurements, the units are different to what they are in Europe and what they are in the UK. So that’s one area. The other one is to do with laundry care. So the laundry care information is either not there, it’s either missing or it’s inconsistent or it’s just incomplete. And that’s right across the range and that’s something that we have to fix.

Miguel Vale

So let us update Smart Advisor with that information. Press it. Again it’s going to the RAG system to a similar search going to LLM updating confidence level, and it has another question but before we go to the this question let me show you a few more features of this of this tool, so there is an “i” where we can press it and break down our information.

So it break down the question. For this case, the question is simple. It just wants to know the impact, right? Customer experience impact. So that’s fine. So we can hide this “i”. And then there is a confidence level down below, where you can click on the confidence level and then you can see what he understands so far. For example, he understands sizing about Europe, UK, America, and so on, and what he’s still lacking. At any moment, you know what the smart advisor knows about us and what he still lacks. And is available to the entire process. But Brian, let’s focus on what he wants now. He wants impact customer experience. So can you tell us a few impact customer experience that we are having?

Rafiq Rajabally

Well, the customer experience is poor, so we have lots of unhappy clients. As I said earlier, there are lots and credits of refunds to try and fix these problems with clients.

And then we’ve got the negative social media that follows, which we’re trying to avoid.

Miguel Vale

Okay, let’s add that information to the agent. So let’s send. And again, it will do the same process again and send. Now, it has a recommendation, but before we process the recommendation, let me just go to the bottom. You see now the confidence level increased to high. This is why it’s giving us the confidence, the recommendation. And also, it still have the rationale here.

But let’s focus on the recommendations. So basically there are three main recommendations. He gave us five, but the main three ones is we really need a centralized data governance framework. We need to implement the data quality management process and the standardization of product information is so important.

And what we should not do, manual processing. So this manual processing sometimes it feels that it’s a good thing, but it just creates problems on the line. And the second thing that we should not do is about our teams. Our teams, they need to collaborate and they need to be in the same page. It’s so important. And finally, our process needs to have a proper testing and validation. So we cannot just go and update the system without the processing. And if we do that, Ryan, we will improve the customer satisfaction. Our operational efficiency goes to the roof and the brand reputation is increased automatically. So what do you think about it?

Rafiq Rajabally

It sounds good. And as you said earlier, this is a self-service capability which presents considerable savings over and above getting a consultant in to tell you that.

Miguel Vale

Yeah, okay, thanks.

Rafiq Rajabally

Fantastic. So here we are, scenario theory, business outcomes. So we knew what the target was, was to improve customer satisfaction across the board.

Rafiq Rajabally

By being able to target a strategy, by doing some diagnosis, by then having almost been able to prescribe what actions to take, we’ve been able to enhance the operational efficiency. The net result of the cumulative effect of the first two points allows us to strengthen brand reputation. And that’s the thing that is going to help us grow the business in the long term.

Rafiq Rajabally

Miguel, I think there’s a comment you want to add to this.

Miguel Vale

Yeah, so basically we saw a smart agent on the data management, but in a way we can create a smart agent in another realm, for example, agenetic architecture or responsible AI or any other topic that we don’t dominate. So these smart agents are there for that, for give us really guide us on implementing that in our organization.

Miguel Vale

So final thoughts on the Onepoint AI ecosystem, Rafiq.

Rafiq Rajabally

Okay, so we’re heading towards the wrap-up and this is what our thoughts are. I think firstly, let’s not forget that we’ve got industry expertise across manufacturing, aerospace, retail, property management and energy. I’d like you to think about our AI agents being the Swiss Army knife equivalent of a multi-tool. So we can provide you a comprehensive end-to-end service, helping you everything from the initial assessment to a strategy to final delivery and beyond.

And don’t make the mistake of thinking that this is only for the bigger, more complex problems. We can work on something quite restricted in scope and we can scale up with you to address other challenges. So I think that’s a key thing. Also, we very much focus on rapid time to value. That is critical for us because otherwise there’s no point in doing this. We need to be able to demonstrate to you rapidly how this can actually promote positive business outcomes for your business. And we have ways of being able to measure the success of this as opposed to maybe a manual or non-automated process.

And finally, it’s not just what we have off the shelf. We also work with bespoke solutions. We have the Onepoint CoLab for that, and we can work with you to tailor a solution to a specific requirement that you may have. Miguel?

Miguel Vale

Thank you, Rafiq. So in the booming marketplace, if you go to the Boomi Marketplace and if you search “buildt by Onepoint”, you’ll see more than 40 recipes and accelerators.

All the accelerators, they have as well a two-minute video that can overview our AI agents and not just the demo itself, but also the architecture of the AI agents. So I really strongly recommend you to go to the Boomi marketplace, search Built by Onepoint and explore what we have there.

If you need any questions, you can send an email to rafiq@onepointltd.com or to myself, miguel@onepointltd.com. If you have anything connected to the Boomi Marketplace, you can send an email to linda.lam@boomi.com. Linda, back to you.

Linda Lam

All right. Thank you both. Again, thank you, Miguel. Yes, if you’re interested in learning in terms of how to become a tech partner, or you would like to build and publish AI agents onto our marketplace, please visit our Technology Partner Program web page, or feel free to reach out to me directly.

All right, so let’s jump into Q&A section here. So this is a question coming from the attendee in terms of, I think it’s regards to the SAP agent. Could we connect with Boomi SAP to S4HANA to issue order directly? And that’s for Miguel.

Miguel Vale

So if you can connect to Boomi SAP, yeah, we have, in fact, our database supports SAP HANA, and it supports SAP HANA, and also supports similar search on SAP HANA. So this is one of the things that we, is one of the database that we support. So a good question. So you can do RAG system using HANA. Okay.

Linda Lam

Let’s see another question here now they’re trickling in um in scenario three uh it shows what to do can i think i or you give examples in terms of how you do it with pools and other people’s advice.

Miguel Vale

Okay so uh i have a good news we have uh two smart agents in the Boomi marketplace one is an accelerator with a more polished version you have a video two five minutes video there you can see. But we have also recipe with a more lightweight of a smart agent that uses a HR database. The database is not polished as the accelerator, but at least you can see how it works. You can download that recipe, you can install, you can change, you can polish. Great question.

Linda Lam

Are there any existing Boomi agent connected to Banner for student ERP? I think that’s a question probably for our side in regards to what’s available with the marketplace. Something definitely we’ll get back to you on. I do advise to go check to see in our margin agent studio to see that that’s available as well for Boomi agents.

Linda Lam

Let’s see, sorry, questions are coming in. I think we have time for one more question. The last question I’m going to ask here is, in terms of building your AI agents, do you prefer to build in-house in terms of an AI expertise, or do you rather use partners to build agents?

Rafiq Rajabally

I’ll give that to Rafiq. Rafiq, you want to answer that, or do you prefer me to answer that?

Rafiq Rajabally

No, I can take that. I just wanted to understand the context of the question because we are a partner. So, you know, I think the value add that we bring is that we understand, A, the AI ecosystem with Boomi that we have. And secondly, you know, we’ve got expert knowledge of the Boomi platform so hopefully that gives a sort of a time to market advantage does that answer the question or Linda.

Linda Lam

Yeah thanks Rafiq. I’m gonna do one more question here before we sum this up what are your consideration about AI agent and security and MCP.

Miguel Vale

So the MCP protocol, the initial version doesn’t consider security, but the agent-to-agent protocol does. I think it’s an evolution. Everyone is now playing with MCPs. There are some values on the MCPs and basically it’s offering tools to agents to be more powerful. Now, I think at the moment everything is experimentation.

But for sure there will be guidelines to use them properly and new versions of the MCP. I don’t know if I fully answered the question.

Rafiq Rajabally

I think you’ve made a good dent in it. I would add that the AI agent space is moving very, very quickly. So there is a lot of evolution in the way that solutions are coming to market. So it is a question of sort of watching this space. And I think there’s a lot of very interesting concepts out there. And, you know, as I suppose people become more confident with the concepts, we’re going to start to see some interesting visions come to take place.

Linda Lam

Awesome. All right, I think we’ll wrap it up with that last question. Thank you so much, Rafiq and Miguel, for presenting today. And thank you, everyone, for joining us. We really hope this presentation shows how you all can achieve business results efficiently and also unlock valuable insights in terms of meaningful data with agents today. Thank you. And we hopefully, for the next time, will be here. Thank you.

Rafiq Rajabally

Thank you.

Miguel Vale

Thank you so much, everyone. Okay. Bye-bye.